SillyTavern

1. Create a detailed character profile using SillyTavern's robust character editor, defining personality, speech patterns, and backstory.

2. Set up a scenario or initial prompt to kickstart the story, optionally adding 'lorebook' entries for world-building.

3. Connect to a powerful LLM backend (local or cloud) known for creative writing and coherence.

4. Interact with the AI character, leveraging SillyTavern's context management to ensure the AI remembers key plot points and character traits, guiding the narrative with user input and AI responses.

1. Configure multiple LLM API connections within SillyTavern, including local models like Oobabooga's web UI and cloud APIs like OpenAI.

2. Create a standard set of prompts or a specific character interaction scenario.

3. Switch between different LLM backends via SillyTavern's interface, applying the same prompt/scenario to each model.

4. Compare the generated responses directly within the same chat interface, noting differences in coherence, creativity, factual accuracy, and token usage for comparative analysis.

1. Develop a 'persona' within SillyTavern that reflects the desired assistant's role (e.g., 'Fantasy Worldbuilder', 'Plot Doctor', 'Character Designer'), complete with instructions and preferred output styles.

2. Feed specific ideas, outlines, or narrative fragments to the AI, leveraging SillyTavern's context window to keep the AI focused on the project's details.

3. Engage in iterative brainstorming sessions, asking the AI to generate plot twists, character dialogues, world details, or scene descriptions, and using the 'regenerate' or 'edit' features to refine outputs.

4. Save conversations and character profiles for future projects, building a library of specialized AI creative partners.